Spaces:
Running
Running
add a few new datasets
Browse files
app.py
CHANGED
@@ -547,6 +547,275 @@ def get_data_ph_eval(eval_mode='zero_shot', fillna=True, rank=True):
|
|
547 |
PH_EVAL_ZERO_SHOT = get_data_ph_eval(eval_mode="zero_shot")
|
548 |
PH_EVAL_FIVE_SHOT = get_data_ph_eval(eval_mode="five_shot")
|
549 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
550 |
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
551 |
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
552 |
|
@@ -792,7 +1061,151 @@ with block:
|
|
792 |
)
|
793 |
|
794 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
795 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
796 |
|
797 |
gr.Markdown(r"""
|
798 |
|
|
|
547 |
PH_EVAL_ZERO_SHOT = get_data_ph_eval(eval_mode="zero_shot")
|
548 |
PH_EVAL_FIVE_SHOT = get_data_ph_eval(eval_mode="five_shot")
|
549 |
|
550 |
+
|
551 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
552 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
553 |
+
|
554 |
+
|
555 |
+
def get_data_sing2eng(eval_mode='zero_shot', fillna=True, rank=True):
|
556 |
+
|
557 |
+
df_list = []
|
558 |
+
|
559 |
+
for model in MODEL_LIST:
|
560 |
+
|
561 |
+
|
562 |
+
results_list = [ALL_RESULTS[model][eval_mode]['sing2eng'][res] for res in ALL_RESULTS[model][eval_mode]['sing2eng']]
|
563 |
+
|
564 |
+
|
565 |
+
try:
|
566 |
+
bleu_score = median([results['bleu_score'] for results in results_list])
|
567 |
+
|
568 |
+
except:
|
569 |
+
print(results_list)
|
570 |
+
bleu_score = -1
|
571 |
+
|
572 |
+
|
573 |
+
res = {
|
574 |
+
"Model Size (Params)": MODEL_TO_SIZE.get(model, ""),
|
575 |
+
"Model": make_clickable_model(model, link=ALL_RESULTS[model]["model_link"]),
|
576 |
+
"BLEU": bleu_score,
|
577 |
+
}
|
578 |
+
|
579 |
+
df_list.append(res)
|
580 |
+
|
581 |
+
|
582 |
+
df = pd.DataFrame(df_list)
|
583 |
+
# If there are any models that are the same, merge them
|
584 |
+
# E.g. if out["Model"] has the same value in two places, merge & take whichever one is not NaN else just take the first one
|
585 |
+
df = df.groupby("Model", as_index=False).first()
|
586 |
+
# Put 'Model' column first
|
587 |
+
#cols = sorted(list(df.columns))
|
588 |
+
cols = list(df.columns)
|
589 |
+
cols.insert(0, cols.pop(cols.index("Model")))
|
590 |
+
df = df[cols]
|
591 |
+
|
592 |
+
if rank:
|
593 |
+
df = add_rank(df, compute_average=True)
|
594 |
+
|
595 |
+
if fillna:
|
596 |
+
df.fillna("", inplace=True)
|
597 |
+
|
598 |
+
return df
|
599 |
+
|
600 |
+
|
601 |
+
SING2ENG_ZERO_SHOT = get_data_sing2eng(eval_mode="zero_shot")
|
602 |
+
SING2ENG_FIVE_SHOT = get_data_sing2eng(eval_mode="five_shot")
|
603 |
+
|
604 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
605 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
606 |
+
|
607 |
+
|
608 |
+
def get_data_flores_ind2eng(eval_mode='zero_shot', fillna=True, rank=True):
|
609 |
+
|
610 |
+
df_list = []
|
611 |
+
|
612 |
+
for model in MODEL_LIST:
|
613 |
+
|
614 |
+
|
615 |
+
results_list = [ALL_RESULTS[model][eval_mode]['flores_ind2eng'][res] for res in ALL_RESULTS[model][eval_mode]['flores_ind2eng']]
|
616 |
+
|
617 |
+
|
618 |
+
try:
|
619 |
+
bleu_score = median([results['bleu_score'] for results in results_list])
|
620 |
+
|
621 |
+
except:
|
622 |
+
print(results_list)
|
623 |
+
bleu_score = -1
|
624 |
+
|
625 |
+
|
626 |
+
res = {
|
627 |
+
"Model Size (Params)": MODEL_TO_SIZE.get(model, ""),
|
628 |
+
"Model": make_clickable_model(model, link=ALL_RESULTS[model]["model_link"]),
|
629 |
+
"BLEU": bleu_score,
|
630 |
+
}
|
631 |
+
|
632 |
+
df_list.append(res)
|
633 |
+
|
634 |
+
|
635 |
+
df = pd.DataFrame(df_list)
|
636 |
+
# If there are any models that are the same, merge them
|
637 |
+
# E.g. if out["Model"] has the same value in two places, merge & take whichever one is not NaN else just take the first one
|
638 |
+
df = df.groupby("Model", as_index=False).first()
|
639 |
+
# Put 'Model' column first
|
640 |
+
#cols = sorted(list(df.columns))
|
641 |
+
cols = list(df.columns)
|
642 |
+
cols.insert(0, cols.pop(cols.index("Model")))
|
643 |
+
df = df[cols]
|
644 |
+
|
645 |
+
if rank:
|
646 |
+
df = add_rank(df, compute_average=True)
|
647 |
+
|
648 |
+
if fillna:
|
649 |
+
df.fillna("", inplace=True)
|
650 |
+
|
651 |
+
return df
|
652 |
+
|
653 |
+
|
654 |
+
FLORES_IND2ENG_ZERO_SHOT = get_data_flores_ind2eng(eval_mode="zero_shot")
|
655 |
+
FLORES_IND2ENG_FIVE_SHOT = get_data_flores_ind2eng(eval_mode="five_shot")
|
656 |
+
|
657 |
+
|
658 |
+
|
659 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
660 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
661 |
+
|
662 |
+
|
663 |
+
def get_data_flores_vie2eng(eval_mode='zero_shot', fillna=True, rank=True):
|
664 |
+
|
665 |
+
df_list = []
|
666 |
+
|
667 |
+
for model in MODEL_LIST:
|
668 |
+
|
669 |
+
|
670 |
+
results_list = [ALL_RESULTS[model][eval_mode]['flores_vie2eng'][res] for res in ALL_RESULTS[model][eval_mode]['flores_vie2eng']]
|
671 |
+
|
672 |
+
|
673 |
+
try:
|
674 |
+
bleu_score = median([results['bleu_score'] for results in results_list])
|
675 |
+
|
676 |
+
except:
|
677 |
+
print(results_list)
|
678 |
+
bleu_score = -1
|
679 |
+
|
680 |
+
|
681 |
+
res = {
|
682 |
+
"Model Size (Params)": MODEL_TO_SIZE.get(model, ""),
|
683 |
+
"Model": make_clickable_model(model, link=ALL_RESULTS[model]["model_link"]),
|
684 |
+
"BLEU": bleu_score,
|
685 |
+
}
|
686 |
+
|
687 |
+
df_list.append(res)
|
688 |
+
|
689 |
+
|
690 |
+
df = pd.DataFrame(df_list)
|
691 |
+
# If there are any models that are the same, merge them
|
692 |
+
# E.g. if out["Model"] has the same value in two places, merge & take whichever one is not NaN else just take the first one
|
693 |
+
df = df.groupby("Model", as_index=False).first()
|
694 |
+
# Put 'Model' column first
|
695 |
+
#cols = sorted(list(df.columns))
|
696 |
+
cols = list(df.columns)
|
697 |
+
cols.insert(0, cols.pop(cols.index("Model")))
|
698 |
+
df = df[cols]
|
699 |
+
|
700 |
+
if rank:
|
701 |
+
df = add_rank(df, compute_average=True)
|
702 |
+
|
703 |
+
if fillna:
|
704 |
+
df.fillna("", inplace=True)
|
705 |
+
|
706 |
+
return df
|
707 |
+
|
708 |
+
|
709 |
+
FLORES_VIE2ENG_ZERO_SHOT = get_data_flores_vie2eng(eval_mode="zero_shot")
|
710 |
+
FLORES_VIE2ENG_FIVE_SHOT = get_data_flores_vie2eng(eval_mode="five_shot")
|
711 |
+
|
712 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
713 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
714 |
+
|
715 |
+
|
716 |
+
def get_data_flores_zho2eng(eval_mode='zero_shot', fillna=True, rank=True):
|
717 |
+
|
718 |
+
df_list = []
|
719 |
+
|
720 |
+
for model in MODEL_LIST:
|
721 |
+
|
722 |
+
|
723 |
+
results_list = [ALL_RESULTS[model][eval_mode]['flores_zho2eng'][res] for res in ALL_RESULTS[model][eval_mode]['flores_zho2eng']]
|
724 |
+
|
725 |
+
|
726 |
+
try:
|
727 |
+
bleu_score = median([results['bleu_score'] for results in results_list])
|
728 |
+
|
729 |
+
except:
|
730 |
+
print(results_list)
|
731 |
+
bleu_score = -1
|
732 |
+
|
733 |
+
|
734 |
+
res = {
|
735 |
+
"Model Size (Params)": MODEL_TO_SIZE.get(model, ""),
|
736 |
+
"Model": make_clickable_model(model, link=ALL_RESULTS[model]["model_link"]),
|
737 |
+
"BLEU": bleu_score,
|
738 |
+
}
|
739 |
+
|
740 |
+
df_list.append(res)
|
741 |
+
|
742 |
+
|
743 |
+
df = pd.DataFrame(df_list)
|
744 |
+
# If there are any models that are the same, merge them
|
745 |
+
# E.g. if out["Model"] has the same value in two places, merge & take whichever one is not NaN else just take the first one
|
746 |
+
df = df.groupby("Model", as_index=False).first()
|
747 |
+
# Put 'Model' column first
|
748 |
+
#cols = sorted(list(df.columns))
|
749 |
+
cols = list(df.columns)
|
750 |
+
cols.insert(0, cols.pop(cols.index("Model")))
|
751 |
+
df = df[cols]
|
752 |
+
|
753 |
+
if rank:
|
754 |
+
df = add_rank(df, compute_average=True)
|
755 |
+
|
756 |
+
if fillna:
|
757 |
+
df.fillna("", inplace=True)
|
758 |
+
|
759 |
+
return df
|
760 |
+
|
761 |
+
|
762 |
+
FLORES_ZHO2ENG_ZERO_SHOT = get_data_flores_zho2eng(eval_mode="zero_shot")
|
763 |
+
FLORES_ZHO2ENG_FIVE_SHOT = get_data_flores_zho2eng(eval_mode="five_shot")
|
764 |
+
|
765 |
+
|
766 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
767 |
+
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
768 |
+
|
769 |
+
|
770 |
+
def get_data_flores_zsm2eng(eval_mode='zero_shot', fillna=True, rank=True):
|
771 |
+
|
772 |
+
df_list = []
|
773 |
+
|
774 |
+
for model in MODEL_LIST:
|
775 |
+
|
776 |
+
|
777 |
+
results_list = [ALL_RESULTS[model][eval_mode]['flores_zsm2eng'][res] for res in ALL_RESULTS[model][eval_mode]['flores_zsm2eng']]
|
778 |
+
|
779 |
+
|
780 |
+
try:
|
781 |
+
bleu_score = median([results['bleu_score'] for results in results_list])
|
782 |
+
|
783 |
+
except:
|
784 |
+
print(results_list)
|
785 |
+
bleu_score = -1
|
786 |
+
|
787 |
+
|
788 |
+
res = {
|
789 |
+
"Model Size (Params)": MODEL_TO_SIZE.get(model, ""),
|
790 |
+
"Model": make_clickable_model(model, link=ALL_RESULTS[model]["model_link"]),
|
791 |
+
"BLEU": bleu_score,
|
792 |
+
}
|
793 |
+
|
794 |
+
df_list.append(res)
|
795 |
+
|
796 |
+
|
797 |
+
df = pd.DataFrame(df_list)
|
798 |
+
# If there are any models that are the same, merge them
|
799 |
+
# E.g. if out["Model"] has the same value in two places, merge & take whichever one is not NaN else just take the first one
|
800 |
+
df = df.groupby("Model", as_index=False).first()
|
801 |
+
# Put 'Model' column first
|
802 |
+
#cols = sorted(list(df.columns))
|
803 |
+
cols = list(df.columns)
|
804 |
+
cols.insert(0, cols.pop(cols.index("Model")))
|
805 |
+
df = df[cols]
|
806 |
+
|
807 |
+
if rank:
|
808 |
+
df = add_rank(df, compute_average=True)
|
809 |
+
|
810 |
+
if fillna:
|
811 |
+
df.fillna("", inplace=True)
|
812 |
+
|
813 |
+
return df
|
814 |
+
|
815 |
+
|
816 |
+
FLORES_ZSM2ENG_ZERO_SHOT = get_data_flores_zho2eng(eval_mode="zero_shot")
|
817 |
+
FLORES_ZSM2ENG_FIVE_SHOT = get_data_flores_zho2eng(eval_mode="five_shot")
|
818 |
+
|
819 |
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
820 |
# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
|
821 |
|
|
|
1061 |
)
|
1062 |
|
1063 |
|
1064 |
+
# dataset 7:
|
1065 |
+
with gr.TabItem("Singlish to English Translation"):
|
1066 |
+
with gr.Row():
|
1067 |
+
gr.Markdown("""
|
1068 |
+
**SING2ENG Leaderboard** 🔮
|
1069 |
+
|
1070 |
+
- **Metric:** BLEU Avg.
|
1071 |
+
- **Languages:** English
|
1072 |
+
""")
|
1073 |
+
|
1074 |
+
with gr.TabItem("zero_shot"):
|
1075 |
+
with gr.TabItem("Overall"):
|
1076 |
+
with gr.Row():
|
1077 |
+
gr.components.Dataframe(
|
1078 |
+
SING2ENG_ZERO_SHOT,
|
1079 |
+
datatype=["number", "markdown"] + ["number"] * len(SING2ENG_ZERO_SHOT.columns),
|
1080 |
+
type="pandas",
|
1081 |
+
)
|
1082 |
+
|
1083 |
+
with gr.TabItem("five_shot"):
|
1084 |
+
with gr.TabItem("Overall"):
|
1085 |
+
with gr.Row():
|
1086 |
+
gr.components.Dataframe(
|
1087 |
+
SING2ENG_FIVE_SHOT,
|
1088 |
+
datatype=["number", "markdown"] + ["number"] * len(SING2ENG_FIVE_SHOT.columns),
|
1089 |
+
type="pandas",
|
1090 |
+
)
|
1091 |
+
|
1092 |
|
1093 |
+
# dataset 8:
|
1094 |
+
with gr.TabItem("FLORES Indonesian to English Translation"):
|
1095 |
+
with gr.Row():
|
1096 |
+
gr.Markdown("""
|
1097 |
+
**flores_ind2eng Leaderboard** 🔮
|
1098 |
+
|
1099 |
+
- **Metric:** BLEU Avg.
|
1100 |
+
- **Languages:** English
|
1101 |
+
""")
|
1102 |
+
|
1103 |
+
with gr.TabItem("zero_shot"):
|
1104 |
+
with gr.TabItem("Overall"):
|
1105 |
+
with gr.Row():
|
1106 |
+
gr.components.Dataframe(
|
1107 |
+
FLORES_IND2ENG_ZERO_SHOT,
|
1108 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_IND2ENG_ZERO_SHOT.columns),
|
1109 |
+
type="pandas",
|
1110 |
+
)
|
1111 |
+
|
1112 |
+
with gr.TabItem("five_shot"):
|
1113 |
+
with gr.TabItem("Overall"):
|
1114 |
+
with gr.Row():
|
1115 |
+
gr.components.Dataframe(
|
1116 |
+
FLORES_IND2ENG_FIVE_SHOT,
|
1117 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_IND2ENG_FIVE_SHOT.columns),
|
1118 |
+
type="pandas",
|
1119 |
+
)
|
1120 |
+
|
1121 |
+
|
1122 |
+
# dataset 9:
|
1123 |
+
with gr.TabItem("FLORES Vitenamese to English Translation"):
|
1124 |
+
with gr.Row():
|
1125 |
+
gr.Markdown("""
|
1126 |
+
**flores_vie2eng Leaderboard** 🔮
|
1127 |
+
|
1128 |
+
- **Metric:** BLEU Avg.
|
1129 |
+
- **Languages:** English
|
1130 |
+
""")
|
1131 |
+
|
1132 |
+
with gr.TabItem("zero_shot"):
|
1133 |
+
with gr.TabItem("Overall"):
|
1134 |
+
with gr.Row():
|
1135 |
+
gr.components.Dataframe(
|
1136 |
+
FLORES_VIE2ENG_ZERO_SHOT,
|
1137 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_VIE2ENG_ZERO_SHOT.columns),
|
1138 |
+
type="pandas",
|
1139 |
+
)
|
1140 |
+
|
1141 |
+
with gr.TabItem("five_shot"):
|
1142 |
+
with gr.TabItem("Overall"):
|
1143 |
+
with gr.Row():
|
1144 |
+
gr.components.Dataframe(
|
1145 |
+
FLORES_VIE2ENG_FIVE_SHOT,
|
1146 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_VIE2ENG_FIVE_SHOT.columns),
|
1147 |
+
type="pandas",
|
1148 |
+
)
|
1149 |
+
|
1150 |
+
|
1151 |
+
|
1152 |
+
# dataset 10:
|
1153 |
+
with gr.TabItem("FLORES Chinese to English Translation"):
|
1154 |
+
with gr.Row():
|
1155 |
+
gr.Markdown("""
|
1156 |
+
**flores_zho2eng Leaderboard** 🔮
|
1157 |
+
|
1158 |
+
- **Metric:** BLEU Avg.
|
1159 |
+
- **Languages:** English
|
1160 |
+
""")
|
1161 |
+
|
1162 |
+
with gr.TabItem("zero_shot"):
|
1163 |
+
with gr.TabItem("Overall"):
|
1164 |
+
with gr.Row():
|
1165 |
+
gr.components.Dataframe(
|
1166 |
+
FLORES_ZHO2ENG_ZERO_SHOT,
|
1167 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_ZHO2ENG_ZERO_SHOT.columns),
|
1168 |
+
type="pandas",
|
1169 |
+
)
|
1170 |
+
|
1171 |
+
with gr.TabItem("five_shot"):
|
1172 |
+
with gr.TabItem("Overall"):
|
1173 |
+
with gr.Row():
|
1174 |
+
gr.components.Dataframe(
|
1175 |
+
FLORES_ZHO2ENG_FIVE_SHOT,
|
1176 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_ZHO2ENG_FIVE_SHOT.columns),
|
1177 |
+
type="pandas",
|
1178 |
+
)
|
1179 |
+
|
1180 |
+
|
1181 |
+
|
1182 |
+
# dataset 10:
|
1183 |
+
with gr.TabItem("FLORES Malay to English Translation"):
|
1184 |
+
with gr.Row():
|
1185 |
+
gr.Markdown("""
|
1186 |
+
**flores_zsm2eng Leaderboard** 🔮
|
1187 |
+
|
1188 |
+
- **Metric:** BLEU Avg.
|
1189 |
+
- **Languages:** English
|
1190 |
+
""")
|
1191 |
+
|
1192 |
+
with gr.TabItem("zero_shot"):
|
1193 |
+
with gr.TabItem("Overall"):
|
1194 |
+
with gr.Row():
|
1195 |
+
gr.components.Dataframe(
|
1196 |
+
FLORES_ZSM2ENG_ZERO_SHOT,
|
1197 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_ZSM2ENG_ZERO_SHOT.columns),
|
1198 |
+
type="pandas",
|
1199 |
+
)
|
1200 |
+
|
1201 |
+
with gr.TabItem("five_shot"):
|
1202 |
+
with gr.TabItem("Overall"):
|
1203 |
+
with gr.Row():
|
1204 |
+
gr.components.Dataframe(
|
1205 |
+
FLORES_ZSM2ENG_FIVE_SHOT,
|
1206 |
+
datatype=["number", "markdown"] + ["number"] * len(FLORES_ZSM2ENG_FIVE_SHOT.columns),
|
1207 |
+
type="pandas",
|
1208 |
+
)
|
1209 |
|
1210 |
gr.Markdown(r"""
|
1211 |
|